Judge orders disclosure of Clearview AI role in DC facial recognition arrest

As law enforcement agencies ramp up their use of facial recognition technology to identify suspects, calls for transparency are getting louder. In Washington, D.C., a judge has ordered prosecutors to hand over more information on facial recognition software provided by Clearview AI, after the biometric tech was used to identify and arrest a shooting suspect on July 1.
Marquis Foster, 43, is charged with assault with intent to kill, for his involvement in a non-fatal shooting that wounded one person on June 8 near Howard University. A report from DC Witness says Foster was identified using surveillance footage from the crime scene; his face was then fed into Clearview’s system, which has a database of 70 billion publicly available facial images.
“The image of the potential suspect that was identified by the software program was then imported into a Metropolitan Police Department (MPD) mugshot’s database, where police identified Foster from an existing image.” The victim then identified Foster from a photo line up.
The reaction from Foster’s lawyer, Elizabeth Weller, is telling: “I’ve never heard of this before,” she told a hearing. While it seems unlikely that Weller is unaware of police use of facial recognition, the process – matching an image with Clearview AI then with a police database, or some combination of such – is the kind of new policing capability that is raising questions about probable cause, mass surveillance and false matches.
The prosecution has pledged to ask the police for more information, and the judge overseeing the case, D.C. Superior Court Judge Neal Kravitz, has said they must provide it to the court and to Foster’s defense. The parties are scheduled to reconvene on July 22 to address the matter further.
New Jersey case omits source code from required info
The question of just how much information a facial recognition provider is obligated to provide during criminal investigations has already reached U.S. courts. In June, the New Jersey Supreme Court unanimously ruled that law enforcement in the state must disclose its use of facial recognition during investigations as part of the discovery process in court proceedings. That includes performance metrics, the name and developer of biometric software in use, reference databases, all versions of the probe image and any match candidates returned.
However, the decision excludes proprietary information such as the source code of the facial recognition software.
Article Topics
biometric matching | biometrics | Clearview AI | facial recognition | law enforcement





Comments